Her doctor's AI scribe put psychedelic mushrooms in her medical record — she’d never taken them

Rebecca Green entered her first appointment with a urologist expecting a routine visit. She willingly agreed to have an AI system record their conversation. Little did she know, that trust would be misplaced. After undergoing surgery for kidney stones in March, Green discovered a false entry in her medical record stating she microdosed psychedelic mushrooms.

The letter to her family doctor claimed this could explain previous kidney bleeding. Green, shocked and in tears, expressed her disbelief, noting she had never taken such substances. The mistake could potentially impact her workplace injury compensation case. Feeling unheard and concerned about potential consequences, Green demanded an explanation.

The urologist apologized and corrected the record, attributing the error to an AI transcription or dictation mishap. However, Green remains worried about the broader implications. She believes human oversight is crucial, as AI systems can make errors that might affect patients' health, employment, or insurance. Despite the excitement surrounding AI scribes, which saw a 40% increase in usage among Australian general practitioners, the technology also presents risks.

AI scribes can omit information, attribute comments incorrectly, or introduce details that weren't mentioned. Inaccurate records can lead to difficulties in correcting or removing the information. Furthermore, patients may not fully grasp the implications of being recorded by AI, potentially altering their willingness to disclose sensitive information.

Consent processes for AI use remain unclear in many clinics, leaving patients uncertain about their rights. While some AI scribe systems perform well, others show flaws, including the inclusion of unmentioned details. As AI becomes more prevalent, ensuring patient privacy and informed consent becomes paramount.